Hacker News Reader

Thanks for your thoughts.

> 1. Even assuming both person A and person B use the service within a week of each other, the odds that they both use the same hastag are astronomically low. Person A says '#wannaGoOut', but person B says '#IWantToGoOut'. There are way too many possible permutations of a single thought.

My hope is that hashtags would become standardized, and hints (https://aytwit.com/thoughter#hints) will also help with this.

> 3. Without an app, I think you'll have a hard time gaining critical mass. If someone has a thought and all it takes to get it out onto Thoughter is to (i) open the app, (ii) type, (iii) and send, then it will be way more likely to be used.

Totally agree. An app would also allow the service to avoid email, and allow encryption on the client-side without the server seeing any plain text at all. Definitely a goal.

> 5. What if I see a cute girl at a coffee store, we hit it off, but I don't have her email address - or even her name? What then?

Good question! I don't have an answer unfortunately. Guess you gotta do things the old-fashioned way. :)

Sell this to Snapchat

This really does fit perfectly with Snapchat’s MO of plausible deniability in messaging; and could be exactly what they need to gain an edge against Instagram.

Seconding talking to Snap.

How would I go about that? I'm just a clueless engineer. What little I know of how the world works, is that they would just steal the idea if it were really attractive to them, and maybe give me a free T-shirt. Am I too jaded?

A Snap engineer could theoretically see this post and steal the idea. It’s usually useless to go around trying to guard an idea. However, if you develop the service enough so that it would be inefficient to rebuild it completely from scratch, then they might pay you for the concept and the technology, which is real IP and can’t be ripped off quite so easily.

I'm fairly certain you'd need a user base before Snap were even remotely interested, or at least prove the idea.

Build an app, get it out there, see whats up

The alternative is they do that and don’t give you the T-shirt.

You should consider a Progressive Web Application: requires SSL, and you can place this on the Apple and Google app stores. You just have to do a small and fair amount of JS witchcraft.

To get the most out of a native app, I would want to do client-side encryption/hashing (right now the server sees plaintext for a millisecond, just like most servers treat passwords), and use mobile notifications instead of email, for the communication mechanism. Not saying a PWA wouldn't fit the bill, but unfortunately it's a bit beyond just wrapping it in a WebView and packaging as an app. I suppose as an MVP maybe? Just go get on the stores?

You can do client side encryption with javacript. Not sure if a PWA can do native notifications though.

Would you be open to piggybacking on existing encrypted messengers, like Telegram and Signal? Maybe even WhatsApp (not sure how their API works). You could also build bots on Matrix.

Yes! This is in the future plans section: https://aytwit.com/about#technical_details__future_plans

I agree email is so five minutes ago, but it was the easiest path to get started.

1. The service will send an email to person A saying "Someone wants to #wannaGoOut". As long as person B doesn't make a typo when copying the hash tag, this problem is solved.

2. Yeah, it's a nod to certain eras of web design, which I like, but other people won't care so I agree.

3. It's email-based in order to make it work for recipients who don't have the app.

5. In 2019, it's probably best to support Facebook, Snapchat, and Instagram as recipient addresses... somehow. Although most of those aren't technically possible. Those platforms aren't really great for building services on top of.

> 1. The service will send an email to person A saying "Someone wants to #wannaGoOut". As long as person B doesn't make a typo when copying the hash tag, this problem is solved.

That's only one of four possible settings, and using it breaks half of the examples: it works great for #reunite, not so great for #breakup

My thoughts too on the breakup one, unless both partners have multiple relationships.

I built this service using Instagram before

Link?

> What if I see a cute girl at a coffee store, we hit it off, but I don't have her email address - or even her name? What then?

Something like Berkanan, a physical proximity-based messaging app that uses Bluetooth, may help.

https://berkanan.chat/

or try talking to her.

anxiety intensifies

Some people do not have enough confidence to risk a public rejection.

Then those people should work on building that confidence.

Psssh, what is this, the 2000s?

This sounds like an awesome idea! Too bad it's not available for Android. But that's how I found out about Firechat.

Machine learning could probably be used to fuzzy-match the contents (probably removing the need for the hashtag altogether). Of course, false-positives could cause problems.

Levenshtein distance, an n-gram comparison, or similar might be good enough instead of having to invite a machine learning solution.